What Is Expected Goals (xG)? A Plain-English Guide for Football Fans
Why the Scoreline Doesn't Tell the Whole Story
Picture a midweek league game that finishes 1-0. The winners scored in the 12th minute and spent the rest of the night chasing shadows. They took four shots all match; the losers took eighteen, hit the post twice, and forced a string of saves. Nine of those eighteen attempts came from inside the penalty area, while three of the winners' four were struck hopefully from 25 yards.
The table records a win, and nothing about the performance. That gap between what happened and what the shots deserved is exactly where frustration lives for fans who watched the game and then read a bare scoreline. A result tells you who finished the job on the night. It says nothing about which side created the better chances, or whether the same performance would produce the same result next week.
Expected goals was built to fill that gap. It puts a number on chance quality, so a 1-0 can be read alongside the shots that produced it. To understand how, you need to know what the metric is actually counting.
What xG Measures, Shot by Shot
Image: DSC2816, by hasgeek, CC BY 2.0
Every shot carries a number between 0 and 1: the share of the time that shot, taken from that spot, with that foot, under that pressure, ends up in the net. Multiply by 100 and you have a percentage. A penalty sits around 0.76, because roughly three in four are converted. A speculative effort from 30 yards is closer to 0.03 or 0.05. Most shots fall somewhere in between, and a tap-in from six yards out with an open goal is worth more than ten hopeful headers.
Team xG is nothing more than those values added together. If a side takes twelve shots worth 0.08 each, their total is 0.96, which reads as "about one goal's worth of chances." The sum doesn't describe any single moment; it describes the volume and quality of everything a team created.
One caveat: Opta, StatsBomb and Understat each build their own model, so the same match can show slightly different totals depending on where you look. The direction of the numbers rarely changes, only the decimals. Knowing what they mean is the first step; the second is finding them on a scoreboard while a match is still going.
How to Read xG on a Live Scoreboard
Image: Web Metrics Grow Up, by TopRankMarketing, CC BY 2.0
Take a full-time panel reading Team A 1-0 Team B, shots 9-14, xG 0.8 to 1.9. The score says Team A won. The shot count says Team B were busier. The xG column says Team B's chances were worth roughly two goals' worth of quality while Team A's were worth under one. Put together, the picture is a side that defended a narrow lead and got the result against a side that created the better openings and did not take them.
That is a description of chances, not a verdict on who deserved what. xG has no opinion about merit, and one match is a small sample: a deflection, a red card, a goalkeeper's best night of the season can swing a single game's numbers well away from the underlying truth. Treat the full-time figures as a record of what happened in front of goal, not a prediction of what comes next.
Most live scoreboards place xG in the same row as shots and possession, so the comparison takes seconds.
Three numbers worth checking
- Compare the two xG totals first: they show which side created the better chances across ninety minutes, whatever the scoreline says.
- Divide xG by shot count to see the shape of a team's threat: many low-value attempts from distance, or fewer clear openings from good positions.
- Subtract xG from goals scored. A positive gap means a side finished above expectation on the day; a negative one flags chances wasted or a keeper kept busy.
Spotting Over- and Under-Performing Teams Over a Season
Image: DSC2816, by hasgeek, CC BY 2.0
That scatter plot is a season-long picture, and it is worth being honest about why the same chart built on two matches would be close to useless. Finishing is noisy over small samples. A side can put four past an opponent from 1.2 xG: a deflection off a defender's shin, a keeper spilling a routine catch, a striker in one of those spells where everything he hits finds a corner. None of that repeats reliably. Widen the window to thirty-eight games and goals tend to drift back toward xG, because the deflections even out and the hot streaks cool. So a team sitting several goals above its xG in November is a reasonable candidate for regression, while a team creating plenty but scoring little is more likely unlucky than bad. The caution is that xG treats every shot of a given type as equal, ignoring goalkeeper quality and shot placement, so the metric narrows the story without finishing it.
The practical habit is to stop looking at one result and start tracking a rolling window. Take a team's last eight to ten matches and note the xG difference across them: chances created minus chances conceded. That figure moves slowly enough to mean something. Compare xG for and xG against together rather than goals alone, because a side can post a healthy xG for while giving up more at the other end, and the two numbers read very differently in combination than either does on its own. Use it to test a manager's tactics directly. If a coach has switched to a back three and the xG against has dropped over ten games, the change is doing work. If the xG for has climbed while results lag, the chances are arriving and the finishing is not. xG now sits on broadcast graphics and most scoreboard apps, so none of this requires a subscription. Treat it as a description of what happened, not a forecast.
What xG Can't Tell You
A shot from twelve yards out wide in the box counts the same whether it was curled into the top corner or rolled straight at the keeper. The model records where it was struck from, not where it was aimed, and it has no idea whether the man in nets is a seasoned international or a debutant. Game state slips past it too: a side 2-0 up at the hour may stop taking risks, so a flat xG for the last thirty minutes reflects intent as much as ability. Providers also build their models differently, weighting distance, angle, foot and pressure in their own proportions, which is why two apps can show different totals for the same match. Compare like with like. And beyond the shots a team concedes, xG says nothing about how the back four held its shape or who tracked the runner. Use it as one lens alongside the eye test and the table, never as a verdict. Treat it as a starting point for watching more closely, asking why the chances came and who was marking whom.